Physical machines
Hosts provide computing hardware with different CPU, GPU, memory and storage capabilities. A machine's actual configuration determines what it can support.
THE SOLARCOMPUTE NETWORK
The distributed data centre starts at home.
SolarCompute is building a network that brings computing hardware, software orchestration and renewable energy infrastructure together. Learn how machines join, how workloads are matched, and where the network is heading.
An overview of network design and development direction — not a live status dashboard.
THE FOUNDATION
Built from connected machines, not one central facility.
The network begins with physical computers running SolarCompute's node software. As the platform develops, suitable machines can contribute processing capacity from residential, commercial and other supported environments.
Node agent
SolarCompute's agent discovers hardware characteristics, connects machines to the platform and reports operational availability for scheduling.
Central orchestration
The platform coordinates eligible machines and routes supported jobs according to approved hardware and workload requirements.
HOW A JOB MOVES
Compatibility before execution.
SolarCompute's Hardware Matrix is designed to approve workloads for specific hardware configurations. A machine does not become eligible simply because it meets a broad minimum specification.
1
Discover
The node agent identifies the machine's hardware profile.
2
Match
The platform checks that profile against the Hardware Matrix.
3
Approve
An approved workload and execution recipe establish what may run.
4
Execute
A compatible available machine runs the assigned job and returns its result.
One exact hardware configuration. One approved workload assignment.
WHAT THE NETWORK SUPPORTS
Different workloads. Different hardware.
SolarCompute is developing a shared platform for a range of compute tasks. Availability depends on supported recipes, hardware compatibility and current capacity.
AI inference
Running supported AI models to produce responses and other outputs.
AI training
Compute-intensive training and related workloads on suitably approved hardware.
Rendering
GPU-accelerated image generation and rendering workflows.
CPU jobs
General-purpose processing suited to approved CPU configurations.
Simulation
Supported computational and simulation tasks matched to appropriate machines.
Custom containers
Controlled execution environments using approved workload recipes.
ENERGY & INFRASTRUCTURE
Designed to work alongside distributed renewable energy.
SolarCompute's long-term vision connects computing with rooftop solar, battery storage and host-side energy management. The aim is to make better use of distributed infrastructure while giving hosts more visibility over their own machines and energy use.
Energy supply, emissions and cost outcomes depend on each site and its operating conditions. Renewable integration and network-wide optimisation will develop in stages.
THE ROAD AHEAD
How the network can evolve.
The current development focus is on validating workload execution and refining hardware compatibility, scheduling and host integration. The following directions are planned areas of expansion, not promises of deployment dates or scale.
Broader host participation
Enable additional compatible machines and host locations to join through a repeatable onboarding and approval process.
More capable workload routing
Refine scheduling, queue management and approved workload recipes as the range of hardware grows.
On-demand resilience
Explore supplementary on-demand compute capacity when suitable distributed machines are busy or unavailable, without requiring permanently idle fallback infrastructure.
Energy-aware operations
Develop closer coordination between available computing capacity, host energy systems and local operating conditions.
PART OF THE NETWORK
A new model for distributed computing.
Discover how to contribute hardware or use supported compute workloads as SolarCompute develops.